Cybersecurity for Industry 4.0 and Advanced Manufacturing Environments with Ensemble Intelligence

Abstract

Traditional cybersecurity architectures incorporate security mechanisms that provide services such as confidentiality, authenticity, integrity, access control, and non-repudiation. These mechanisms are used extensively to prevent computer and network intrusions and attacks. For instance, access control services prevent unauthorized access to cyber resources such as computers, networks, and data. However, the modern Internet security landscape is characterized by attacks that are voluminous, constantly evolving, extremely fast, persistent, and highly sophisticated Schnackenberg et al. (2000), Anuar et al. (2010). These characteristics impose significant challenges on preventive security services. Consequently, methodologies that enable autonomic detection and response to cyberattacks should be employed synergistically with prevention techniques in order to achieve effective defense-in-depth strategies and robust cybersecurity systems. This is especially true for the critical systems belonging to Industry 4.0 systems. In this chapter, we describe how we have integrated cyberattack detection and response mechanisms into our Software-Defined Cloud Manufacturing architecture. The cyberattack detection algorithm described in this chapter is based on ensemble intelligence with neural networks whose outputs are fed into a neuro-evolved neural network oracle. The oracle produces an optimized classification output that is used to provide feedback to active attack response mechanisms within our software-defined cloud manufacturing system. The underlying goal of this chapter is to show how computational intelligence approaches can be used to defend critical Industry 4.0 systems as well as other Internet-driven systems.